Under　the　application　background　of　network　security　evaluation,　a　mechanism　for　situation　element　extraction　based　on　Particle　Swarm　Optimization　(PSO)　and　Fuzzy　Neural　Network　(FNN)　is　proposed.　Firstly,　the　input　dataset　of　historical　situation　element　is　pre-fuzzed　and　then　transformed　into　fuzzy　logic　rule　which　can　be　mapped　between　neural　network　layers.　Meanwhile,　PSO　is　used　to　achieve　global　optimization　of　BP　network＇s　weight　value　and　threshold,　and　then　an　extraction　model　based　on　FNN　and　PSO　(PSO-FNN)　is　built　up.　Experiment　results　prove　that　this　extraction　mechanism　is　effective　in　situation　element　extraction,　and　can　be　applied　in　the　area　of　situation　extraction　for　network　security　situation　awareness.　©　2008　IEEE.